Abstract
Neonates rely on their innate immune system, and neutrophils in particular, to recognize and combat life-threatening bacterial infections. Pretreatment with lipopolysaccharide (LPS), a toll-like receptor (TLR) 4 agonist, improves survival to polymicrobial sepsis in neonatal mice by enhancing neutrophil recruitment. To understand the response of human neonatal neutrophils to TLR4 stimulation, ex vivo spontaneous neutrophil migration, neutrophil transcriptomics, and cytokine production in the presence and absence of LPS was measured directly from whole blood of adults, term neonates, and preterm neonates. Spontaneous neutrophil migration was measured on novel microfluidic devices with time-lapse imaging for ten hours. Genome-wide neutrophil transcriptomics and plasma cytokine concentrations were also determined. Preterm neonates had significantly fewer spontaneously migrating neutrophils at baseline, and both term and preterm neonates had decreased neutrophil velocity, compared to adults. In the presence of LPS stimulation, the number of spontaneously migrating neutrophils was reduced in preterm neonates compared to term neonates and adults. Neutrophil velocity was not significantly different among groups with LPS stimulation. Preterm neonates upregulated expression of genes associated with the recruitment and response of neutrophils following LPS stimulation, but failed to upregulate the expression of genes associated with antimicrobial and antiviral responses. Plasma levels of IL-1β, IL-6, IL-8, MIP-1α, and TNF-α increased in response to LPS stimulation in all groups, but IL-10 was increased only in term and preterm neonates. In conclusion, age-specific changes in spontaneous neutrophil migration counts are not affected by LPS despite changes in gene expression and cytokine production.
Keywords: immunology, innate immunity, inflammation, microfluidics, neutrophil migration, lipopolysaccharide
1. Introduction
Sepsis is a major cause of morbidity and mortality in neonates [1–3]. A crucial component of the neonatal response to sepsis is pathogen recognition via various innate immune cells including neutrophils [4]. We have previously demonstrated preterm neonates have diminished neutrophil chemotaxis to the bacterial chemotactic peptide, N-formylmethionine-leucyl-phenylalanine (fMLP) [5]. This finding may contribute to the increased risk of sepsis and mortality among the most vulnerable neonatal population. In neonatal mice, we have demonstrated that in vivo pretreatment with the toll-like receptor (TLR) 4 agonist lipopolysaccharide (LPS) improves survival to polymicrobial sepsis by enhancing neutrophil recruitment [6]. We also discovered unique cytokine and neutrophil transcriptomic profiles in umbilical cord blood compared to healthy young adult peripheral blood following ex vivo LPS stimulation [7]. In addition, LPS at a concentration of 1000 ng/ml has been shown to be a potent stimulus for inducing random migration in human adult neutrophils [8].
Whole blood studies are limited in the neonatal population given concerns about the volume of blood that would be required to perform a study such as this, especially among extremely low birth weight (ELBW, birth weight <1000g) neonates for whom total circulating blood volumes may be less than 100 milliliters [9]. We have overcome this challenge with the use of a novel microfluidic technology which can precisely measure neutrophil migration counts and velocities from as little as two microliters of whole blood [10]. By testing neutrophil function in the presence of the subject’s own plasma, this technology mirrors a physiological environment and allows detailed comparisons of neutrophil motility in a matter of hours. In addition, these techniques do not require extensive neutrophil isolation procedures which may induce artificial stimulation [11]. Therefore, we aimed to determine the impact of TLR4 stimulation on whole blood from adults, term and preterm neonates by measuring spontaneous neutrophil migration, neutrophil transcriptomics, and cytokine production. We hypothesized that ex vivo exposure to LPS would improve spontaneous neutrophil migration and this could be explained by changes in neutrophil transcription. Surprisingly, ex vivo LPS exposure had only minimal effects on spontaneous neutrophil migration velocity and no effect on the number of migrating neutrophils. Transcriptional analysis confirmed these findings with no changes in transcription of genes involved in the cell movement, migration, or homing of neutrophils.
2. Materials and Methods
2.1. Study design
This prospective observational study was conducted between September 2016 and August 2017 at UF Health Shands Children’s Hospital, a 202-bed tertiary medical center including a Level IV Neonatal Intensive Care Unit (NICU). The study was approved by the Institutional Review Board prior to initiation. All neonates admitted to the Newborn Nursery and NICU were screened for inclusion. Neonates born between 24 and 42 weeks gestational age were eligible for participation. Neonates were excluded from enrollment if any major congenital anomaly was known or suspected. Neonates were withdrawn from the study if the sample collection did not coincide temporally with a clinically indicated blood draw. Written informed consent was obtained from the parents or guardians of neonatal subjects prior to sample or data collection. Consent was provided only to review the neonatal medical record. Healthy adults aged 21 – 45 years were also recruited for participation. Adult subjects were excluded from enrollment if any of the following was reported: pre-existing organ dysfunction, positive for human immunodeficiency virus, recent use of oncolytics, current use of steroids, or history of autoimmune disease. Written informed consent was obtained from all adult subjects prior to sample collection.
2.2. Blood sample collection
Whole blood samples were collected in tubes anticoagulated with lithium heparin (Becton, Dickinson, and Company, Franklin Lakes, NJ, USA) from young adults (21–45 years old), term neonates (gestational age 37–42 weeks), and preterm neonates (gestational age 24–36 weeks). For adults, a four milliliter blood sample was collected by venipuncture. For neonates, a single 700 microliter blood sample was collected at the time of a clinically indicated blood draw. Blood was collected from term neonates in the Newborn Nursery prior to discharge between 24 and 36 hours of life and at the same time as the state’s newborn screening panel. Blood was collected for preterm neonates in the NICU on day four of life in order to reduce the impact of potentially immunomodulating medications that were administered to the mother prior to delivery and to allow the family time to consider study participation.
2.3. Patient data collection
Neonatal data including demographics, clinical variables, and outcomes were recorded from the electronic medical records of the birth hospitalization. Prolonged premature rupture of membranes (PROM) was defined as any rupture of membranes that persisted for more than 24 hours and occurred prior to the onset of labor. Apgar scores were recorded at one and five minutes following birth. Neonatal antibiotics, corticosteroids, and indomethacin were recorded if administered any time prior to study sample collection. Use of mechanical ventilation was recorded based on the respiratory status of the neonate at the time of study sample collection.
2.4. Spontaneous neutrophil migration assay
Microfluidic devices with a single whole blood loading chamber (WBLC), erythrocyte filters, migration channels, and posts were utilized as previously described and validated [10] (Fig. 1a). Erythrocyte filters block the dispersion of erythrocytes into the migration channel, whereas spontaneously migrating neutrophils can travel through without a decrease in speed. Erythrocyte filters additionally prevent the entrance of other, non-neutrophil leukocytes into the migration channels with neutrophils consisting of greater than 96% of cells entering the channels. Posts located at the end of the migration channels allow neutrophils to change direction and continue spontaneous movement. Microfluidic spontaneous migration devices were filled with either Hank’s Balanced Salt Solution (HBSS) with bovine albumin [final concentration 0.2 mg/ml] (Sigma-Aldrich, St. Louis, MO, USA) alone or a mix of HBSS with albumin and lipopolysaccharide from Escherichia coli 026:B6 [final concentration 1000 ng/ml] (Sigma-Aldrich, St. Louis, MO, USA). The device was then placed in a desiccator under vacuum for 10 minutes to ensure complete filling of the device. Following device filling, two microliters of whole blood stained with Hoechst Stain Solution (Sigma-Aldrich, St. Louis, MO, USA) were pipetted into the WBLC. Time-lapse imaging was performed every 30 seconds for 10 hours on a Nikon Bio Station IMq™ microscope inside of a biochamber maintained at 37° C and 5% CO2. Neutrophil migration counts were calculated using Nikon Imaging Software Elements™ High Content Analysis (Nikon Instruments Inc., Melville, NY, USA). Neutrophil velocities were calculated using ImageJ software (NIH, Bethesda, MD, USA) [12].
Figure 1. Spontaneous neutrophil migration assay.
a. Microfluidic device. Neutrophils migrate from the center whole blood loading chamber (WBLC), past the erythrocyte filters, and into the migration channels that are sized specifically for neutrophils but not other cell populations. Once in the migration channels, neutrophils are able to spontaneously migrate. Posts located at the end of the migration channels allow neutrophils to change direction and continue spontaneous movement. Individual neutrophils are tracked and neutrophil migration count and velocity are obtained with time-lapse imaging. b. Unstimulated neutrophil migration. Average spontaneous neutrophil migration profiles over 10 hours for unstimulated adults (n=10), unstimulated term neonates (n=10), and unstimulated preterm neonates (n=10). Unstimulated preterm neonates had a reduction in number of spontaneous migrating neutrophils compared to unstimulated term neonates (p=0.0034) and adults (p=0.0280). c. Unstimulated neutrophil velocity. At baseline, both term and preterm neonates had a reduction in spontaneous migrating neutrophil velocity compared to adults (unstimulated preterm 13.57, unstimulated term 15.33, unstimulated adult 21.48 μm/min; p=0.0007). ** indicates values that are significantly different between unstimulated term neonates and adults, (p<0.01). *** indicates values that are significantly different between unstimulated preterm neonates and adults, (p<0.001). Lines represent means with standard deviation. d. Stimulated neutrophil migration. Average spontaneous neutrophil migration profiles over 10 hours for stimulated adults (n=10), stimulated term neonates (n=10), and stimulated preterm neonates (n=10). Stimulated preterm neonates had a reduction in number of spontaneous migrating neutrophils compared to stimulated term neonates (p=0.0332) and adults (p=0.0151). e. Stimulated neutrophil velocity. There was no significant difference in neutrophil velocity among stimulated groups (stimulated preterm 17.47, stimulated term 17.06, stimulated adult 19.18 μm/min; p=0.3607). Lines represent means with standard deviation.
2.5. Neutrophil transcriptome assay
Whole blood samples were aliquoted and mixed 1:1 with HBSS with bovine albumin [final concentration 0.2 mg/ml] alone or HBSS with albumin and LPS from Escherichia coli 026:B6 [final concentration 100 ng/ml] (Sigma-Aldrich, St. Louis, MO, USA). Aliquots were then incubated on a rocker for two hours at 37° C and 5% CO2. Following incubation, neutrophils were captured from whole blood using positive selection with anti-CD66b monoclonal antibody, clone 80H3 (BioRad, Hercules, CA, USA) coated microfluidic devices as previously described and validated [5, 7, 11, 13]. The microfluidic devices have greater than 95% neutrophil purity and capture a fixed number of neutrophils, thus the total number of neutrophils captured were equivalent among groups regardless of any differences in total or differential counts. Captured neutrophils were lysed en bloc using RLT buffer (Qiagen, Valencia, CA, USA). The cell lysate was passed through a QiaShredder column, and RNA was extracted from lysates using QIAGEN RNeasy Mini Kit (Qiagen, Valencia, CA, USA). RNA integrity was assessed with an Agilent 2100 Bioanalyzer (Santa Clara, CA, USA). RNA was labeled using Nugen Ovation Pico WTA System V2 (NuGEN Technologies, Inc, San Carlos, CA, USA). Resulting cDNA was labeled and fragmented using Nugen Encore Biotin Module, hybridized onto GeneChip Human Transcriptome Array 2.0 (Affymetrix, Santa Clara, CA, USA) and processed following manufacturer’s instructions. Genome-wide expression was normalized using RMA as implemented in Partek Genomics Suite, Version 6.6 (Partek Inc, St. Louis, Mo).
2.6. Genomic analysis
Genome-wide expression was compared among six groups (unstimulated and stimulated adults, unstimulated and stimulated term neonates, unstimulated and stimulated preterm neonates) using both unsupervised and supervised analyses. In the unsupervised analysis, probe sets with a coefficient of variation (CoV) greater than 50% were identified. In the supervised analysis, significant probe sets were identified using an f-test at p<0.001. Genome-wide expression was also compared among unstimulated groups and stimulated groups separately, using an f-test at p<0.001. Finally, individual groups (adults, term neonates, preterm neonates) were compared with and without stimulation using a paired t-test at p<0.001. Leave-one-out cross-validation was used to compute the correct classification rate for each sample. The classification rate was compared to chance based on 1000 random permutations. Genomic analysis was performed using BRB ArrayTools™ Version:4.5.1-Stable Release. Hierarchical clustering was generated with DNA-Chip Analyzer (dChip) software. Principal Component Analysis (PCA) was generated with Partek® Genomics Suite® Version:6.6. Once significant probe sets were identified, Gene Ontology™ analysis was conducted as implemented in BRB ArrayTools™. Fold changes of significant genes were calculated between stimulated and unstimulated groups.
2.7. Functional pathway analysis
Functional pathway analysis was performed on probe sets that were differentially expressed between groups based on the f-test at p<0.001 and had greater than a 1.5 fold-change using Ingenuity® Pathway Analysis (IPA®) software (Ingenuity Systems, Redwood City, CA, USA). IPA® predicts canonical pathways and biological functions based on the expression values of gene groups and prior knowledge of expected effects of the gene products. Predictions with a negative log p-value greater than 1.3 and a z-score greater than 2 or less than −2 were considered significant.
2.8. Cytokine production assay
Whole blood samples were aliquoted and mixed 1:1 with HBSS with bovine albumin [final concentration 0.2 mg/ml] alone or HBSS with albumin and LPS from Escherichia coli 026:B6 [final concentration 100 ng/ml] (Sigma-Aldrich, St. Louis, MO, USA). Aliquots were then incubated on a rocker for two hours at 37° C and 5% CO2. Following incubation, whole blood was centrifuged at 1800 × g for 10 minutes at 4° C, and cell-free supernatant was collected and stored at −80° C until batch processing to reduce the degree of intra-assay variability. The concentrations of interleukin (IL)-1β, IL-6, IL-8 (CXCL8), IL-10, IL-12 p70, interferon-γ (IFNγ), tumor necrosis factor-α (TNFα), macrophage inflammatory protein-1α (MIP-1α, CCL3), monocyte chemotactic protein-1 (MCP-1, CCL2), and interferon gamma-induced protein-10 (IP-10, CXCL10) were determined using Milliplex® Multiplex kits (EMD Millipore, Billerica, MA, USA) on the Luminex® MAGpix Multiplex reader. Cytokine assay sensitivities (minimum detectable concentrations, pg/mL) are provided in Supplemental Table 1.
2.9. Statistical analysis
Data are presented as either frequency and percentage, median and range, or mean and standard deviation. Analysis of variance (ANOVA), Student’s t-tests, and Tukey’s honest significant difference (HSD) tests were used to compare continuous parametric variables as appropriate. Wilcoxon tests and Kruskal-Wallis tests were used to compare continuous nonparametric variables. Fischer’s exact test was used for comparison of categorical variables. All significance tests were two-sided, with p-values less than 0.05 considered statistically significant. Statistical analyses were performed with R statistical software package (Vienna, Austria), SAS software (Cary, NC, USA), and Prism GraphPad Software (La Jolla, CA, USA).
3. Results
3.1. Patient data
Patient data regarding demographics, clinical variables, and outcomes are provided in Table 1. Maternal age and rectovaginal Group B Streptococcus (GBS) colonization were similar between mothers of term and preterm neonates. There was no evidence of clinical chorioamnionitis in either group. Mothers of preterm neonates had a greater incidence of prolonged PROM and were more likely to receive antepartum antibiotics, corticosteroids, and magnesium. Preterm neonates were more likely to be born via cesarean section compared to term neonates. Presence and duration of labor could not be accurately extracted from the neonatal medical record and access to the maternal medical record was not granted. As expected, preterm neonates had lower gestational age, birth weight, and Apgar scores at one and five minutes compared to term neonates. Preterm neonates were also more likely to receive mechanical ventilation, antibiotics, and indomethacin.
Table 1.
Patient Demographics, Clinical Variables, and Outcomes
| Maternal and pregnancy characteristics | Term (n=22) | Preterm (n=21) | p value |
|---|---|---|---|
| Maternal age, median (range) | 29 (16–36) | 30 (18–39) | 0.7132 |
| GBS Colonization, n (%) | |||
| Positive | 7 (32) | 8 (38) | 0.7546 |
| Negative | 14 (64) | 10 (48) | 0.3640 |
| Unknown | 1 (5) | 3 (14) | 0.3449 |
| Prolonged PROM, n (%) | 0 (0) | 7 (33) | 0.0036 |
| Clinical Chorioamnionitis, n (%) | 0 (0) | 0 (0) | 1.0000 |
| Antepartum antibiotics, n (%) | 6 (27) | 14 (67) | 0.0148 |
| Antepartum corticosteroids, n (%) | 0 (0) | 13 (62) | <0.0001 |
| Antepartum magnesium, n (%) | 1 (5) | 15 (71) | <0.0001 |
| Cesarean section, n (%) | 7 (32) | 16 (76) | 0.0058 |
| Neonatal characteristics | Term (n=22) | Preterm (n=21) | p value |
| Male, n (%) | 11 (50) | 11 (52) | 1.0000 |
| Gestational age, median (range) | 39.8 (37.3–41.7) | 32.3 (25.0–35.7) | <0.0001 |
| Birth weight in grams, median (range) | 3610 (2633–4342) | 1878 (470–3446) | <0.0001 |
| Apgar score at 1 min, median (range) | 8 (2–9) | 5 (1–9) | 0.0004 |
| Apgar score at 5 min, median (range) | 9 (8–9) | 7 (3–9) | <0.0001 |
| Mechanical ventilation, n (%) | 0 (0) | 5 (24) | 0.0211 |
| Neonatal antimicrobial treatment, n (%) | 0 (0) | 15 (71) | <0.0001 |
| Neonatal corticosteroids, n (%) | 0 (0) | 3 (14) | 0.1078 |
| Indomethacin, n (%) | 0 (0) | 6 (29) | 0.0089 |
3.2. Spontaneous neutrophil migration
Unstimulated preterm neonatal neutrophils migrated in significantly fewer numbers compared to unstimulated term neonatal (p=0.0034) or adult (p=0.0280) neutrophils based on a time series random intercept model (Fig. 1b). Spontaneous neutrophil migration velocities were significantly reduced in both unstimulated term and preterm neonates compared to unstimulated adults (unstimulated preterm 13.57, unstimulated term 15.33, unstimulated adult 21.48 μm/min; p=0.0007) (Fig. 1c). The velocity of unstimulated preterm neonatal neutrophils was comparable to that of term neonates. With ex vivo LPS stimulation, preterm neonatal neutrophils migrated in significantly fewer numbers compared to LPS-stimulated term neonatal (p=0.0332) or adult (p=0.0151) neutrophils (Fig. 1d). Whereas unstimulated neutrophil velocities from term and preterm neonates were reduced compared to young adults, there was no significant difference in neutrophil velocities among LPS-stimulated groups (stimulated preterm: 17.47, stimulated term: 17.06, stimulated adult: 19.18 μm/min; p=0.3607) (Fig. 1e).
3.3. Neutrophil transcriptomics
An unsupervised analysis of gene expression from the six groups (unstimulated and stimulated adults, unstimulated and stimulated term, unstimulated and stimulated preterm) identified 4013 probe sets, representing 2594 genes, differentially expressed with a CoV greater than 50%. When these 4013 probe sets were hierarchically clustered, LPS-stimulated groups were distinguishable from unstimulated groups, and adults and neonates further separated into distinct groups. The unsupervised clustering did not distinguish gene expression patterns between unstimulated term and preterm neonates.
Using a supervised analysis, the expression of 4582 probe sets, representing 3977 genes, significantly differed among the six groups (p<0.001). When these 4582 probe sets were hierarchically clustered, the expression pattern first distinguished adults from neonates (Fig. 2a). Adult and neonatal groups then separated based on the presence or absence of LPS stimulation. Leave-one-out cross-validation using a 3-nearest neighbor prediction model was able to correctly predict class label at a rate of 75% compared to the expected rate for chance alone of 17%. Based on 1000 random permutations this result was significant at p<0.001. Thus, the gene expression differences between the groups have predictive validity and are informative as to the groups.
Figure 2. Neutrophil transcriptomics and cytokine production in response to TLR4 stimulation.
a. Supervised analysis hierarchical clustering. Using a supervised analysis, the expression of 3977 genes significantly differed among the six groups (p<0.001). The significant genes were then hierarchically clustered based on the Pearson correlations. The gene expression patterns between unstimulated and stimulated neonates were more similar to each other than to unstimulated and stimulated adults. A: Adults, N: Neonates, U: Unstimulated, S: Stimulated. b. Principal component analysis of unstimulated samples. The expression of 398 genes significantly differed among the three unstimulated groups (p<0.001). Expression patterns from unstimulated adult samples (red) are seen on the left side of the image, while expression patterns from unstimulated term samples (orange) and unstimulated preterm samples (green) are seen on the upper and lower right of the image, respectively. c. Principal component analysis of stimulated samples. The expression of 1779 genes significantly differed among the three stimulated groups (p<0.001). Expression patterns from stimulated adult samples (dark blue) are seen tightly clustered on the left side of the image, while expression patterns from stimulated preterm samples (purple) and stimulated term samples (light blue) are seen on the upper and lower right of the image, respectively. d. Cytokine production. Median concentrations and interquartile ranges of cytokines from unstimulated and stimulated adults (n=12), term neonates (n=13), and preterm neonates (n=14). In response to TLR4 stimulation, adults and neonates significantly increased IL-1β, IL-6, IL-8, MIP-1α, and TNF-α concentrations. IL-10 concentrations were increased by TLR4 stimulation in neonates, while adults demonstrated no significant change. IP-10 concentrations were significantly increased by TLR4 stimulation in adults and term neonates, while preterm neonates had no significant change. IL-12 concentrations were significantly increased by TLR4 stimulation in adults and preterm neonates, while term neonates had no significant change. IFNγ concentrations were significantly increased by TLR4 stimulation in only preterm neonates. MCP-1 concentrations did not significantly change by TLR4 stimulation in any group. * indicates values that are significantly different between unstimulated and stimulated groups, (p<0.05).
Analyses of neutrophil gene expression patterns from unstimulated groups and stimulated groups performed separately demonstrated distinct gene expression patterns of adults, term neonates, and preterm neonates. The expression of 398 genes differed significantly among the unstimulated samples (p<0.001), and the expression of 1779 genes differed significantly among the stimulated samples (p<0.001). Principal component analyses (PCA) demonstrated clear separation among unstimulated and stimulated samples from adults, term neonates, and preterm neonates. (Fig. 2b and 3c).
The number of genes that were differentially expressed in response to LPS stimulation increased with age (preterm: 760, term: 916, adult: 1129). A common cluster of 222 unique genes was identified whose expression changed in stimulated samples as compared to unstimulated samples regardless of age group. Shared upregulated genes included key components of innate immune response as shown in Table 2.
Table 2.
Top 20 shared genes whose expression increased in response to LPS stimulation and associated fold changesa
| Symbol | Name | Adult | Term | Preterm |
|---|---|---|---|---|
| IL23A | interleukin 23 alpha subunit | 74.4 | 29.3 | 16.9 |
| IRG1 | immunoresponsive 1 homolog | 20.5 | 26.6 | 18.5 |
| CXCL10 | chemokine (C-X-C motif) ligand 10 | 46 | 10 | 6.2 |
| IL6 | interleukin 6 | 25.2 | 19.6 | 14.4 |
| CCL20 | chemokine (C-C motif) ligand 20 | 22.1 | 16.6 | 12.8 |
| IL1A | interleukin 1 alpha | 12.2 | 13 | 15.5 |
| SLC39A8 | solute carrier family 39 member 8 | 20.4 | 9.8 | 6.6 |
| ITGB8 | integrin subunit beta 8 | 13.4 | 11.3 | 6.2 |
| AK4 | adenylate kinase 4 | 7.1 | 7.2 | 8.1 |
| SLAMF7 | SLAM family member 7 | 10.5 | 6.3 | 4.6 |
| FLT1 | fms related tyrosine kinase 1 | 6.6 | 6.2 | 7.3 |
| CCL7 | chemokine (C-C motif) ligand 7 | 8.6 | 4.2 | 5.5 |
| CD69 | CD69 molecule | 9.4 | 3.2 | 3.8 |
| PI3 | peptidase inhibitor 3 | 2.6 | 3.9 | 9.7 |
| TNFAIP6 | TNF alpha induced protein 6 | 6 | 5 | 5.1 |
| MX1 | MX dynamin like GTPase 1 | 8.4 | 3.7 | 3.4 |
| GCH1 | GTP cyclohydrolase 1 | 5.9 | 5.3 | 4.2 |
| RIN2 | Ras and Rab interactor 2 | 4.3 | 4.7 | 6.2 |
| CLIC4 | chloride intracellular channel 4 | 7.3 | 4.6 | 3.2 |
| NCR3LG1 | natural killer cell cytotoxicity receptor 3 ligand 1 | 5.7 | 4.8 | 3.6 |
Top genes were sorted by greatest mean fold change across all groups
3.4. Functional predictions
Ingenuity® Pathway Analysis software was used to predict key canonical pathways represented by probe sets that were differentially expressed based on f-test at p<0.001 and had greater than a 1.5 fold-change. In response to LPS stimulation, adults, term neonates, and preterm neonates had significant upregulation of genes representing innate inflammatory response pathways, including activation of interferon-regulator factor (IRF) by cytosolic pattern recognition receptors, inflammasome pathway, and role of retinoic acid-inducible gene I (RIG-I) -like receptors in antiviral innate immunity (Supplemental Table 2). In addition, preterm neonates uniquely had significant upregulation of phosphoinositide 3-kinase/protein kinase B (PI3K/AKT) signaling and triggering receptor expressed on myeloid cells 1 (TREM1) signaling, whereas adults and term neonates did not demonstrate significant changes in these pathways. Adults, term neonates, and preterm neonates had significant downregulation of sphingosine-1-phosphate signaling in response to LPS stimulation.
Biological function predictions in response to LPS stimulation demonstrated adults, term neonates, and preterm neonates had significant overexpression of genes involved in immune response of cells, cell-mediated response, and inflammatory response (Supplemental Table 3). Adults and term neonates also had a significant overexpression of genes involved in antimicrobial and antiviral response, whereas preterm neonates did not. Term and preterm neonates had a significant overexpression of genes involved in the recruitment of neutrophils. Preterm neonates exclusively had a significant overexpression of genes involved in the response of neutrophils. There were no significant differences in expression of genes involved in the cell movement, migration, or homing of neutrophils.
3.5. Cytokine production
In response to LPS stimulation, adults, term neonates, and preterm neonates had a significant increase in IL-1β, IL-6, IL-8, MIP-1α, and TNF-α plasma concentrations (Fig. 2d). IL-10 plasma concentrations were increased following LPS stimulation in term and preterm neonates, while adults demonstrated no significant change. IP-10 concentrations were significantly increased by LPS stimulation in adults and term neonates, while preterm neonates demonstrated no significant change. IL-12 concentrations were significantly increased by LPS stimulation in adults and preterm neonates, while term neonates had no significant change. IFN-γ concentrations were significantly increased by LPS stimulation in only preterm neonates. MCP-1 concentrations did not significantly change by LPS stimulation in any group, although baseline concentrations were significantly higher in term and preterm neonates.
4. Discussion
Previously, we demonstrated preterm human neonates had diminished chemotactic neutrophil migration and velocity [5]. Here, we report for the first time that preterm neonates had a decreased number of spontaneous migrating neutrophils compared to adults and term neonates. Term and preterm neonates also had reduced spontaneous migrating neutrophil velocity compared to adults. In addition, we demonstrate that TLR4 stimulation had no effect on the numbers of spontaneous migrating neutrophils in any age group. However, in the presence of a TLR4 agonist, spontaneous migrating neutrophil velocity was not significantly different among adults, term neonates, and preterm neonates, suggesting partial but incomplete mitigation.
Neutrophil transcriptomic and cytokine profiles were compared among adults, term neonates, and preterm neonates with and without TLR4 stimulation. Transcriptomic profiles were significantly different with and without stimulation. Moreover, the number of genes that differed between adults, term neonates, and preterm neonates was greater with stimulation than without stimulation. This finding suggests that although there are differences in neutrophil transcriptome at baseline as demonstrated here and previously [5], differences in transcriptomics are amplified in a TLR4-stimulated state. This is in agreement with our previous results comparing neutrophil transcriptomics of umbilical cord blood and healthy adults at baseline and after 24 hours of TLR4 stimulation [7]. We also observed a positive correlation between the number of genes that changed in response to stimulation and age. This correlation with age is consistent with findings from our multi-cohort, pooled analysis of 636 adults, children, infants, and neonates with and without clinical sepsis [14].
Functional predictions from the neutrophil transcriptome suggest that in response to TLR4 stimulation, adult and neonatal neutrophils have a significant upregulation of innate inflammatory response pathways. In addition, preterm neonatal neutrophils had significant upregulation of PI3K/AKT signaling. Several murine studies have demonstrated that PI3K plays an important role in neutrophil migration [15–17]. Both adult and neonatal neutrophils had downregulation of sphingosine-1-phosphate signaling which has previously been shown to inhibit TLR-activated CXCL8 secretion [18]. In response to TLR4 stimulation, preterm neonates had no significant change in expression of genes involved in antimicrobial and antiviral response, whereas there was significant overexpression in term neonates and adults. This finding is in agreement with previous studies that report a diminished transcriptomic immune response in neonatal sepsis [14]. Importantly, this is the first time the impact of ex vivo TLR4 stimulation on neutrophil transcriptomics has been reported from whole blood of term and preterm neonates. Following TLR4 stimulation, preterm neonates had a significant overexpression of genes involved in the recruitment and response of neutrophils which was not observed in adults. Somewhat surprisingly, there were no overall differences in the expression of genes involved in the cell movement, migration, or homing of neutrophils, suggesting that the differences are more likely due to variations in the host inflammatory response than in the machinery involved in cell movement.
Whole blood from neonates and adults stimulated with a TLR4 agonist exhibited significant increase in the concentrations of multiple pro-inflammatory cytokines and chemokines including IL-1β, IL-6, IL-8, MIP-1α, and TNF-α. In particular, IL-8 has a key role in the migration, recruitment, and activation of neutrophils [19]. Preterm neonates alone had a significant increase in IFNγ, but an attenuated IP-10 response suggesting that interferon-signaling pathways may be implicated [20]. Cuenca et al. previously showed that TIR-domain-containing adapter protein-inducing interferon-β (TRIF) dependent pathways are differentially regulated in neonatal mice [21]. Term and preterm neonates had a significant increase in the concentrations of the anti-inflammatory cytokine IL-10, while there was no significant change in adults. Conversely, adults had a significant IL-12 response to TLR4 stimulation. Similarly, we previously reported a significant increase in IL-10 levels in umbilical cord blood following 24 hours of ex vivo TLR4 stimulation, which was not observed in peripheral blood obtained from healthy adults [7]. In neonatal mice, IL-10 levels were significantly increased 4 hours following in vivo administration of a TLR4 agonist [6]. Taken together, these finding suggest that adults have a more biased Th1 response than neonates who have a more Th2 inflammatory response.
There are a number of potential limitations that require further discussion. First, the neonatal populations are inherently heterogeneous. As expected, preterm neonates were born at lower birth weights and Apgar scores compared to term neonates and were more likely to receive mechanical ventilation, antibiotics, and indomethacin following birth given their prematurity. These clinical characteristics mirror the vast differences of these two neonatal populations. Since electronic medical records were only reviewed for neonatal subjects, maternal characteristics such as antenatal medication administration are subject to reporting bias in the neonates’ records. This bias is expected to be minimum and affect all neonates similarly. Additionally, based on the neonates’ records, the presence and duration of labor could not be accurately determined. The impact of this on our reported measurements is unknown, however is presumed to be less impactful in this study compared to umbilical cord blood studies. Samples from term neonates were collected between 24 and 36 hours of life because of the current practice of rapid discharge of healthy term newborns. In contrast, samples from preterm neonates were delayed until day four of life. The difference in sample timing was deliberate to reduce the impact of potentially immunomodulating medications that were administered to the mother prior to delivery and to the preterm neonate after birth. In addition, insufficient blood sample volumes prevented measurement of cell counts and differentials in the ex vivo whole blood studies. Therefore, we are unable to determine whether the differences in cytokine production reflect differences in cell number and type, or whether they can be explained by differences in cell functionality. Importantly, both microfluidic devices used are not dependent on the starting number of neutrophils. Despite these potential study limitations, the findings that we have presented accurately reflect the clinical scenario and are in themselves novel and contribute to the literature providing previously unpublished knowledge and understanding of human neonatal immune response particularly in preterm neonates which are lacking due to limited sampling volumes and opportunities to study human preterm neonates.
In conclusion, we demonstrate that a novel microfluidic technology can be utilized in the neonatal population to successfully measure spontaneous neutrophil migration. For the first time, preterm neonates have been shown to have a significant reduction in spontaneous neutrophil migration counts and velocities compared to adults, which along with a decreased ability to follow a chemoattractant gradient to fMLP, may help explain their increased susceptibility to infection. With TLR4 stimulation, spontaneous neutrophil migration counts remained decreased among preterm neonates, but velocity was not significantly different among stimulated groups. Finally, we demonstrated unique, age-specific transcriptomic and cytokine profiles in response to TLR4 stimulation that suggest neonates have an attenuated interferon-dependent and biased Th2 response. Further studies are required to evaluate the efficacy of TLR4 agonist administration to enhance neonatal immunity.
Supplementary Material
Supplemental Table 1: Cytokine assay sensitives (minimum detectable concentrations, pg/mL).
Supplemental Table 2: Predicted canonical pathways among adults, term neonates, and preterm neonates in response to TLR4 stimulation.
Supplemental Table 3: Predicted biological functions and disease pathways among adults, term neonates, and preterm neonates in response to TLR4 stimulation.
The video demonstrates the spontaneous neutrophil migration device in operation. Hoechst stained neutrophils migrate through the erythrocyte filters along the migration channel and then change direction at the posts before continuing to migrate along the channels. Images are recorded every 30 seconds for 10 hours, and the video has been modified to reduce 10 hours of sampling time to two minutes.
Acknowledgments
This work was supported by grants from the National Institute of General Medical Sciences (NIGMS; K08 GM106143, R01 GM097531, and P50 GM111152). SLR was supported by a clinical research training fellowship awarded by the Surgical Infection Society Foundation. RBH and JAS were supported by a NIGMS post-doctoral training grant (T32 GM008721). The authors would like to thank all study participants and nursing staff at UF Health Shands Children’s Hospital who made this study possible.
Footnotes
Conflict of Interest Disclosure
All authors have read the journal’s policy on disclosure of potential conflicts of interest and have none to declare.
Authorship
SLR, JLW, LLM, DI, and SDL developed the hypothesis and designed the experiments. FE and DI provided the microfluidic spontaneous migration devices and technical support for the migration assay. SLR, RBH, TJM, JCR, JAS, MCL, RU, and HVB performed the experiments and analyzed the results. SLR, LLM, and SDL prepared the original manuscript. All authors reviewed and revised the final manuscript. No editorial support was used of the preparation of the manuscript.
References
- 1.Lawn JE, Kerber K, Enweronu-Laryea C, Cousens S. 3.6 million neonatal deaths--what is progressing and what is not? Seminars in perinatology. 2010;34:371–386. doi: 10.1053/j.semperi.2010.09.011. [DOI] [PubMed] [Google Scholar]
- 2.Shane AL, Stoll BJ. Neonatal sepsis: progress towards improved outcomes. The Journal of infection. 2014;68(Suppl 1):S24–32. doi: 10.1016/j.jinf.2013.09.011. [DOI] [PubMed] [Google Scholar]
- 3.Khan AM, Morris SK, Bhutta ZA. Neonatal and Perinatal Infections. Pediatric clinics of North America. 2017;64:785–798. doi: 10.1016/j.pcl.2017.03.008. [DOI] [PubMed] [Google Scholar]
- 4.Raymond SL, Stortz JA, Mira JC, Larson SD, Wynn JL, Moldawer LL. Immunological Defects in Neonatal Sepsis and Potential Therapeutic Approaches. Frontiers in pediatrics. 2017;5:14. doi: 10.3389/fped.2017.00014. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Raymond SL, Mathias BJ, Murphy TJ, Rincon JC, Lopez MC, Ungaro R, Ellett F, Jorgensen J, Wynn JL, Baker HV, et al. Neutrophil chemotaxis and transcriptomics in term and preterm neonates. Translational research: the journal of laboratory and clinical medicine. 2017;190:4–15. doi: 10.1016/j.trsl.2017.08.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Wynn JL, Scumpia PO, Winfield RD, Delano MJ, Kelly-Scumpia K, Barker T, Ungaro R, Levy O, Moldawer LL. Defective innate immunity predisposes murine neonates to poor sepsis outcome but is reversed by TLR agonists. Blood. 2008;112:1750–1758. doi: 10.1182/blood-2008-01-130500. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Mathias B, Mira JC, Rehfuss JP, Rincon JC, Ungaro R, Nacionales DC, Lopez MC, Baker HV, Moldawer LL, Larson SD. LPS Stimulation of Cord Blood Reveals a Newborn-Specific Neutrophil Transcriptomic Response and Cytokine Production. Shock (Augusta, Ga) 2017;47:606–614. doi: 10.1097/shk.0000000000000800. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Aomatsu K, Kato T, Fujita H, Hato F, Oshitani N, Kamata N, Tamura T, Arakawa T, Kitagawa S. Toll-like receptor agonists stimulate human neutrophil migration via activation of mitogen-activated protein kinases. Immunology. 2008;123:171–180. doi: 10.1111/j.1365-2567.2007.02684.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Sisson TR, Whalen LE, Telek A. The blood volume of infants. II. The premature infant during the first year of life. The Journal of pediatrics. 1959;55:430–446. doi: 10.1016/s0022-3476(59)80281-x. [DOI] [PubMed] [Google Scholar]
- 10.Ellett F, Jorgensen J, Marand AL, Liu YM, Martinez MM, Sein V, Butler KL, Lee J, Irimia D. Diagnosis of sepsis from a drop of blood by measurement of spontaneous neutrophil motility in a microfluidic assay. Nature Biomedical Engineering. 2018;2:207–214. doi: 10.1038/s41551-018-0208-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Warner EA, Kotz KT, Ungaro RF, Abouhamze AS, Lopez MC, Cuenca AG, Kelly-Scumpia KM, Moreno C, O’Malley KA, Lanz JD, et al. Microfluidics-based capture of human neutrophils for expression analysis in blood and bronchoalveolar lavage. Laboratory investigation; a journal of technical methods and pathology. 2011;91:1787–1795. doi: 10.1038/labinvest.2011.94. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Arena ET, Rueden CT, Hiner MC, Wang S, Yuan M, Eliceiri KW. Quantitating the cell: turning images into numbers with ImageJ. Wiley interdisciplinary reviews. Developmental biology. 2017:6. doi: 10.1002/wdev.260. [DOI] [PubMed] [Google Scholar]
- 13.Kotz KT, Xiao W, Miller-Graziano C, Qian WJ, Russom A, Warner EA, Moldawer LL, De A, Bankey PE, Petritis BO, et al. Clinical microfluidics for neutrophil genomics and proteomics. Nature medicine. 2010;16:1042–1047. doi: 10.1038/nm.2205. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Raymond SL, Lopez MC, Baker HV, Larson SD, Efron PA, Sweeney TE, Khatri P, Moldawer LL, Wynn JL. Unique transcriptomic response to sepsis is observed among patients of different age groups. PloS one. 2017;12:e0184159. doi: 10.1371/journal.pone.0184159. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Sasaki T, Irie-Sasaki J, Jones RG, Oliveira-dos-Santos AJ, Stanford WL, Bolon B, Wakeham A, Itie A, Bouchard D, Kozieradzki I, et al. Function of PI3Kgamma in thymocyte development, T cell activation, and neutrophil migration. Science (New York, NY) 2000;287:1040–1046. doi: 10.1126/science.287.5455.1040. [DOI] [PubMed] [Google Scholar]
- 16.Li Z, Jiang H, Xie W, Zhang Z, Smrcka AV, Wu D. Roles of PLC-beta2 and -beta3 and PI3Kgamma in chemoattractant-mediated signal transduction. Science (New York, NY) 2000;287:1046–1049. doi: 10.1126/science.287.5455.1046. [DOI] [PubMed] [Google Scholar]
- 17.Hirsch E, Katanaev VL, Garlanda C, Azzolino O, Pirola L, Silengo L, Sozzani S, Mantovani A, Altruda F, Wymann MP. Central role for G protein-coupled phosphoinositide 3-kinase gamma in inflammation. Science (New York, NY) 2000;287:1049–1053. doi: 10.1126/science.287.5455.1049. [DOI] [PubMed] [Google Scholar]
- 18.Sharma N, Akhade AS, Qadri A. Sphingosine-1-phosphate suppresses TLR-induced CXCL8 secretion from human T cells. Journal of leukocyte biology. 2013;93:521–528. doi: 10.1189/jlb.0712328. [DOI] [PubMed] [Google Scholar]
- 19.Hammond ME, Lapointe GR, Feucht PH, Hilt S, Gallegos CA, Gordon CA, Giedlin MA, Mullenbach G, Tekamp-Olson P. IL-8 induces neutrophil chemotaxis predominantly via type I IL-8 receptors. Journal of immunology (Baltimore, Md: 1950) 1995;155:1428–1433. [PubMed] [Google Scholar]
- 20.Ellis TN, Beaman BL. Interferon-gamma activation of polymorphonuclear neutrophil function. Immunology. 2004;112:2–12. doi: 10.1111/j.1365-2567.2004.01849.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Cuenca AG, Joiner DN, Gentile LF, Cuenca AL, Wynn JL, Kelly-Scumpia KM, Scumpia PO, Behrns KE, Efron PA, Nacionales D, et al. TRIF-dependent innate immune activation is critical for survival to neonatal gram-negative sepsis. Journal of immunology (Baltimore, Md: 1950) 2015;194:1169–1177. doi: 10.4049/jimmunol.1302676. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplemental Table 1: Cytokine assay sensitives (minimum detectable concentrations, pg/mL).
Supplemental Table 2: Predicted canonical pathways among adults, term neonates, and preterm neonates in response to TLR4 stimulation.
Supplemental Table 3: Predicted biological functions and disease pathways among adults, term neonates, and preterm neonates in response to TLR4 stimulation.
The video demonstrates the spontaneous neutrophil migration device in operation. Hoechst stained neutrophils migrate through the erythrocyte filters along the migration channel and then change direction at the posts before continuing to migrate along the channels. Images are recorded every 30 seconds for 10 hours, and the video has been modified to reduce 10 hours of sampling time to two minutes.